A Multi-method Approach with Machine Learning to Evaluating the Distribution and Intensity of Prehistoric Land Use in Eastern Iberia
R Markdown Scripts for Reproducing Analyses
Get all the packages ready
Load data
# Load Medland survey collection data
medland_survey2014_2017 <- read_csv("data/medland_survey2014_2017.csv", locale = locale(encoding = "850"))
Rows: 1501 Columns: 28── Column specification ───────────────────────────────────────────────────────────────────────────────
Delimiter: ","
chr (5): study.area, sector, subsector, visibility, time.date
dbl (23): ID, zone, collection, xcoord, ycoord, area.sqm, undiag.lithics, burins, notch.dent, MP.to...
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.Figures 1 and 2 generated in GIS software
Figure 3: Surface Visibility and Artifact Recovery
medland_survey2014_2017 %>%
dplyr::filter(area.sqm>0 & total.lithics>0 & !is.na(visibility)) %>%
mutate(lithic.density=total.lithics/area.sqm) %>%
ggplot(aes(x=lithic.density), xlim=.02) +
geom_histogram(binwidth = .001) +
scale_y_log10() +
scale_x_continuous(limits = c(0,0.01), breaks = c(0,.002, .004, .006, .008)) +
labs(title = "Surface Visibility and Artifact Recovery",
x="lithic artifacts / km^2",
y='count of patches') +
facet_grid(factor(visibility)~study.area) +
theme_bw(base_size = 20) +
theme(axis.title.x = element_markdown())
ANOVA for Figure 3: all survey areas and each survey area
cat("\nCanal de Navarrés survey area\n")
Canal de Navarrés survey areawith(medland_survey2014_2017 %>%
dplyr::filter(total.lithics>0 & !is.na(visibility) & study.area == "Canal de Navarrés") %>%
mutate(lithic.density=total.lithics/area.sqm),
aov(lithic.density~visibility)) %>% summary() Df Sum Sq Mean Sq F value Pr(>F)
visibility 2 0.0000022 1.104e-06 0.451 0.637
Residuals 194 0.0004746 2.446e-06 cat("\nHoya de Buñol survey area\n")
Hoya de Buñol survey areawith(medland_survey2014_2017 %>%
dplyr::filter(total.lithics>0 & !is.na(visibility) & study.area == "Hoya de Buñol") %>%
mutate(lithic.density=total.lithics/area.sqm),
aov(lithic.density~visibility)) %>% summary() Df Sum Sq Mean Sq F value Pr(>F)
visibility 2 1.220e-06 6.095e-07 0.4 0.671
Residuals 98 1.493e-04 1.523e-06 cat("\nCocina/Catadau survey area\n")
Cocina/Catadau survey areawith(medland_survey2014_2017 %>%
dplyr::filter(total.lithics>0 & !is.na(visibility) & study.area == "Cocina/Catadau") %>%
mutate(lithic.density=total.lithics/area.sqm),
aov(lithic.density~visibility)) %>% summary() Df Sum Sq Mean Sq F value Pr(>F)
visibility 2 0.0001894 9.470e-05 2.662 0.0758 .
Residuals 83 0.0029531 3.558e-05
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1Random Forest Model for Chronological Unmixing (Figure 4 and Table 4)
Data preparation
Prepare data for Valencia dated assemblages
# Modify training data based on ML and Bayesian testing:
# Merge ENEOL and MNEOL
# Remove undiagnostic lithics for age estimates
# Sort factor levels chronologically
training_data_E_Iberia <- training_data_E_Iberia %>%
select(-undiag.lithics, -total.lithics, -citation) %>%
mutate(period = replace(period, period == "MNEOL", "ENEOL"),
period = factor(period,
levels = c("MP", "UP", "EPI", "MESO", "ENEOL", "LNEOL")))Prepare data from survey collections for applying Random Forest model
# filter out assemblages with only undiagnostic lithics
medland_survey2014_2017_lithics <- medland_survey2014_2017 %>%
dplyr::filter(undiag.lithics < total.lithics) %>%
select(ID, c(13:27))
# make separate table of assemblage ID and provenience
medland_survey2014_2017_info <- medland_survey2014_2017 %>%
dplyr::filter(undiag.lithics < total.lithics) %>%
mutate(assemblage = paste(study.area, "-", zone, "-", sector, "-", subsector, sep = "")) %>%
select(ID, study.area, assemblage)
# calculate lithic density for each collection patch
medland_survey2014_2017_density <- medland_survey2014_2017 %>%
dplyr::filter(undiag.lithics < total.lithics) %>%
mutate(assemblage = paste(study.area, "-", zone, "-", sector, "-", subsector, sep = ""),
density.km2 = total.lithics*1000/area.sqm)%>%
select(ID, total.lithics, area.sqm,density.km2)Split data into training and test sets
# Partition into training and hold out test / validation sample
set.seed(456) ## if we want to make it completely reproducible
vl.split <- training_data_E_Iberia %>%
rsample::initial_split(., prop=.75)
vl.train <- rsample::training(vl.split)
vl.test <- rsample::testing(vl.split)
# save ID data for later analysis
vl.test.id <- vl.test %>%
select(ID, period)
create v-fold objects for replicable and comparable cross-validation
10 folds using all the training data
set.seed(456)
vf10.all <- vfold_cv(training_data_E_Iberia %>% select(-ID),v=10)10 folds using the 75% training data split
set.seed(456)
vf10.train <- vfold_cv(vl.train %>% select(-ID),v=10)Test Random Forest Model for Estimating Age of Surface Assemblages
Create and evaluate RF model using a 75% split (vl.split)
Define and instantiate a random forest model
valencia.lithics.rf.mod <-
rand_forest(trees=500) %>%
set_engine("ranger", importance = "impurity") %>%
set_mode("classification")
print(valencia.lithics.rf.mod)Random Forest Model Specification (classification)
Main Arguments:
trees = 500
Engine-Specific Arguments:
importance = impurity
Computational engine: ranger Fit the model to the training data
valencia.lithics.rf.fit <-
valencia.lithics.rf.mod %>%
fit(as.factor(period) ~ ., data = vl.train[,2:ncol(vl.train)])
print(valencia.lithics.rf.fit)parsnip model object
Ranger result
Call:
ranger::ranger(x = maybe_data_frame(x), y = y, num.trees = ~500, importance = ~"impurity", num.threads = 1, verbose = FALSE, seed = sample.int(10^5, 1), probability = TRUE)
Type: Probability estimation
Number of trees: 500
Sample size: 96
Number of independent variables: 15
Mtry: 3
Target node size: 10
Variable importance mode: impurity
Splitrule: gini
OOB prediction error (Brier s.): 0.2416893 Optional graph: variable importance for random forest model with training set
valencia.lithics.rf.fit %>%
extract_fit_engine() %>%
vip(aesthetics = list(color = "black", fill = "#26ACB5"), num_features = 15) + theme_minimal()
Extract the fitted data
valencia.lithics.rf.predicted <-
valencia.lithics.rf.fit %>%
predict(vl.test) %>%
bind_cols(vl.test.id[1:2], ., predict(valencia.lithics.rf.fit, vl.test, type="prob")) %>%
rename(predicted.age = .pred_class,
MP=.pred_MP,
UP=.pred_UP,
EPI=.pred_EPI,
MESO=.pred_MESO,
ENEOL=.pred_ENEOL,
LNEOL=.pred_LNEOL,
true.age = period) %>%
mutate(true.age = factor(true.age,
levels=c("MP", "UP", "EPI", "MESO", "ENEOL", "LNEOL")),
predicted.age = factor(predicted.age,
levels=c("MP", "UP", "EPI", "MESO", "ENEOL", "LNEOL")))
print(valencia.lithics.rf.predicted)Figure 4: Graph random forest predictions for test set
valencia.lithics.rf.predicted %>%
pivot_longer(cols = 4:ncol(valencia.lithics.rf.predicted),
names_to = "period", values_to = "probability") %>%
mutate(period = factor(period,
levels = c("MP", "UP", "EPI", "MESO", "ENEOL", "LNEOL"))) %>%
ggplot(aes(x=period, y=probability)) +
geom_line(group=1) +
geom_vline(aes(xintercept = true.age), color="red", size=2, alpha=.5) +
geom_vline(aes(xintercept = predicted.age), color="blue", size=0.8) +
labs(title="Random Forest Predictions for Each Assemblage",
subtitle="black line indicates probability, blue line indicates predicted age, & red line indicates known age",
x="time period",
y="probability of predicted time period") +
facet_wrap(vars(ID), ncol = 7) +
theme_bw(base_size = 20) +
theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1),
strip.text.x = element_text(size = 14))
Create confusion matrix for 75% training/test split
library(caret)
with(valencia.lithics.rf.predicted,
caret::confusionMatrix(true.age, predicted.age))Confusion Matrix and Statistics
Reference
Prediction MP UP EPI MESO ENEOL LNEOL
MP 13 0 0 0 0 0
UP 0 2 0 0 2 0
EPI 0 1 1 1 0 0
MESO 0 0 0 3 1 0
ENEOL 0 0 0 1 5 0
LNEOL 0 0 0 0 2 1
Overall Statistics
Accuracy : 0.7576
95% CI : (0.5774, 0.8891)
No Information Rate : 0.3939
P-Value [Acc > NIR] : 2.407e-05
Kappa : 0.6788
Mcnemar's Test P-Value : NA
Statistics by Class:
Class: MP Class: UP Class: EPI Class: MESO Class: ENEOL Class: LNEOL
Sensitivity 1.0000 0.66667 1.00000 0.60000 0.5000 1.00000
Specificity 1.0000 0.93333 0.93750 0.96429 0.9565 0.93750
Pos Pred Value 1.0000 0.50000 0.33333 0.75000 0.8333 0.33333
Neg Pred Value 1.0000 0.96552 1.00000 0.93103 0.8148 1.00000
Prevalence 0.3939 0.09091 0.03030 0.15152 0.3030 0.03030
Detection Rate 0.3939 0.06061 0.03030 0.09091 0.1515 0.03030
Detection Prevalence 0.3939 0.12121 0.09091 0.12121 0.1818 0.09091
Balanced Accuracy 1.0000 0.80000 0.96875 0.78214 0.7283 0.96875Create and evaluate cross-validated random forest model using 10 folds
Define and instantiate a random forest model workflow and fit to cross-validated data set
#Create workflow step
valencia.lithics.rf.wf <-
workflow() %>%
add_model(valencia.lithics.rf.mod) %>%
add_formula(period ~ .) #The predictor is contained in add_formula method
set.seed(456) # For reproducibility
valencia.lithics.rf.xv.fit <-
valencia.lithics.rf.wf %>%
fit_resamples(vf10.all,
control=control_resamples(save_pred = TRUE))! Fold01: internal: No observations were detected in `truth` for level(s): 'UP', 'LNEOL'
Computation will...
! Fold02: internal: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL'
Computation wil...
! Fold03: internal: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL'
Computation wil...
! Fold04: internal: No observations were detected in `truth` for level(s): 'UP'
Computation will proceed ...
! Fold06: internal: No observations were detected in `truth` for level(s): 'EPI', 'MESO'
Computation will...
! Fold09: internal: No observations were detected in `truth` for level(s): 'LNEOL'
Computation will proce...
! Fold10: internal: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL'
Computation wil...print(valencia.lithics.rf.xv.fit)# Resampling results
# 10-fold cross-validation
There were issues with some computations:
- Warning(s) x3: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL' Computation wi... - Warning(s) x1: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL' Computation wi... - Warning(s) x1: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL' Computation wi... - Warning(s) x1: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL' Computation wi... - Warning(s) x1: No observations were detected in `truth` for level(s): 'EPI', 'LNEOL' Computation wi...
Use `collect_notes(object)` for more information.Collect the cross-validated metrics
# Collect the metrics using another model with cross-validation
valencia.lithics.rf.xv.meanpreds <- tune::collect_metrics(valencia.lithics.rf.xv.fit)
print(valencia.lithics.rf.xv.meanpreds)Optional Graph: variable importance for cross-validated random forest model
valencia.lithics.rf.mod %>%
fit(as.factor(period) ~ ., data = training_data_E_Iberia[,2:ncol(training_data_E_Iberia)]) %>%
extract_fit_engine() %>%
vip(aesthetics = list(color = "black", fill = "#26ACB5"), num_features = 15) +
theme_minimal()
Extract predictions of random forest model with cross-validation
valencia.lithics.rf.xv.predictions <- collect_predictions(valencia.lithics.rf.xv.fit, summarize = TRUE) %>%
arrange(.row) %>%
rename(predicted.age = .pred_class,
MP=.pred_MP,
UP=.pred_UP,
EPI=.pred_EPI,
MESO=.pred_MESO,
ENEOL=.pred_ENEOL,
LNEOL=.pred_LNEOL,
true.age=period) %>%
select(-.row, -.config) %>%
relocate(predicted.age, .after = true.age) %>%
bind_cols(training_data_E_Iberia[1], .)
print(valencia.lithics.rf.xv.predictions)Optional Graph: results for cross-validated random forest predictions
valencia.lithics.rf.xv.predictions %>%
pivot_longer(cols = 4:ncol(valencia.lithics.rf.xv.predictions), names_to = "period", values_to = "probability") %>%
mutate(period = factor(period, levels = c("MP","UP","EPI","MESO","ENEOL","MNEOL","LNEOL"))) %>%
ggplot(aes(x=period, y=probability)) +
geom_line(group=1) +
geom_vline(aes(xintercept = true.age), color="red", size=2, alpha=.5) +
geom_vline(aes(xintercept = predicted.age), color="blue", size=0.8) +
labs(title="Random Forest with Cross-Validated Predictions",
subtitle="Time Periods for Each Assemblage",
x="predicted time period\n(blue line indicates prediction & red line indicates radiocarbon age)") +
facet_wrap(vars(ID)) +
theme_bw(base_size = 16) +
theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1))
Table 4: Confusion matrix for cross-validated random forest predictions
caret::confusionMatrix(
as.factor(valencia.lithics.rf.xv.predictions$true.age),
as.factor(valencia.lithics.rf.xv.predictions$predicted.age))Confusion Matrix and Statistics
Reference
Prediction MP UP EPI MESO ENEOL LNEOL
MP 51 0 0 0 0 0
UP 0 9 1 0 5 0
EPI 1 3 2 0 0 0
MESO 0 3 0 11 5 0
ENEOL 0 1 0 6 22 0
LNEOL 0 0 0 0 3 6
Overall Statistics
Accuracy : 0.7829
95% CI : (0.7018, 0.8507)
No Information Rate : 0.4031
P-Value [Acc > NIR] : < 2.2e-16
Kappa : 0.7073
Mcnemar's Test P-Value : NA
Statistics by Class:
Class: MP Class: UP Class: EPI Class: MESO Class: ENEOL Class: LNEOL
Sensitivity 0.9808 0.56250 0.66667 0.64706 0.6286 1.00000
Specificity 1.0000 0.94690 0.96825 0.92857 0.9255 0.97561
Pos Pred Value 1.0000 0.60000 0.33333 0.57895 0.7586 0.66667
Neg Pred Value 0.9872 0.93860 0.99187 0.94545 0.8700 1.00000
Prevalence 0.4031 0.12403 0.02326 0.13178 0.2713 0.04651
Detection Rate 0.3953 0.06977 0.01550 0.08527 0.1705 0.04651
Detection Prevalence 0.3953 0.11628 0.04651 0.14729 0.2248 0.06977
Balanced Accuracy 0.9904 0.75470 0.81746 0.78782 0.7771 0.98780Create random forest model from all training data to estimate ages of surface collections from survey
Define and instantiate a Random Forest model using all the known training data
medland.survey.rf.mod <-
rand_forest(trees=500) %>%
set_engine("ranger") %>%
set_mode("classification")
print(medland.survey.rf.mod)Random Forest Model Specification (classification)
Main Arguments:
trees = 500
Computational engine: ranger Extract fit of Random Forest model
medland.survey.rf.fit <-
medland.survey.rf.mod %>%
fit(as.factor(period) ~ ., data = training_data_E_Iberia[,2:ncol(training_data_E_Iberia)])
print(medland.survey.rf.fit)parsnip model object
Ranger result
Call:
ranger::ranger(x = maybe_data_frame(x), y = y, num.trees = ~500, num.threads = 1, verbose = FALSE, seed = sample.int(10^5, 1), probability = TRUE)
Type: Probability estimation
Number of trees: 500
Sample size: 129
Number of independent variables: 15
Mtry: 3
Target node size: 10
Variable importance mode: none
Splitrule: gini
OOB prediction error (Brier s.): 0.2341059 Apply Random Forest model to surface collections and generate age estimate predictions for each collection
medland.survey.rf.predicted <-
medland.survey.rf.fit %>%
predict(medland_survey2014_2017_lithics) %>%
bind_cols(medland_survey2014_2017_info, .,
predict(medland.survey.rf.fit, medland_survey2014_2017_lithics, type="prob")) %>%
rename(predicted.period = .pred_class,
MP=.pred_MP,
UP=.pred_UP,
EPI=.pred_EPI,
MESO=.pred_MESO,
ENEOL=.pred_ENEOL,
LNEOL=.pred_LNEOL) %>%
mutate(predicted.period = factor(predicted.period,
levels=c("MP", "UP", "EPI", "MESO", "ENEOL", "LNEOL")))
print(medland.survey.rf.predicted)Calculate age probabilities to assign to collections with only undiagostic lithics
Use the 10th percentile of probabilities from random forest model with dated collections
valencia.lithics.rf.probquantiles <- with(valencia.lithics.rf.xv.predictions %>%
pivot_longer(cols = 4:ncol(valencia.lithics.rf.xv.predictions),
values_to = "probability"),
quantile(probability, probs = c(.1, .25, .5, .75)))
print(paste("10th percentile of random forest probabilities =",
valencia.lithics.rf.probquantiles[1]))[1] "10th percentile of random forest probabilities = 0.00173557343664358"Optional Graph: show quantiles for random forest probabilities
# show this in a graph
valencia.lithics.rf.xv.predictions %>%
pivot_longer(cols = 4:ncol(valencia.lithics.rf.xv.predictions), names_to = "period", values_to = "probability") %>%
mutate(period = factor(period, levels = c("MP","UP","EPI","MESO","ENEOL","MNEOL","LNEOL"))) %>%
ggplot(aes(x=probability)) +
geom_density() +
geom_vline(xintercept = valencia.lithics.rf.probquantiles[1], color = 'red') +
geom_vline(xintercept = valencia.lithics.rf.probquantiles[2], color = 'green') +
geom_vline(xintercept = valencia.lithics.rf.probquantiles[3], color = 'blue') +
geom_vline(xintercept = valencia.lithics.rf.probquantiles[4], color = 'green') +
labs(title="Distribution of Combined Random Forest Predictions",
x="age prediction probabilities \n(blue = median, red = 10th percentile, green = 25th and 75th percentile")
Create files for additional graphing and for output for mapping
# Create base file for graphing and output
medland.survey.rf.graph <-
left_join(select(medland_survey2014_2017, ID, study.area, zone, sector, subsector, total.lithics, area.sqm) %>%
mutate(assemblage = paste(study.area, "-", zone, "-", sector, "-", subsector, sep = "")),
medland.survey.rf.predicted) %>%
dplyr::filter(total.lithics>0) %>%
mutate(density.km2 = total.lithics/area.sqm/1000)Joining, by = c("ID", "study.area", "assemblage")# Create output file
# Use 10th percentile for overall ubiquity for patches with only undiagnostic lithics
medland.survey.rf.out <- medland.survey.rf.graph %>%
mutate(MP = replace_na(MP, valencia.lithics.rf.probquantiles[1]),
UP = replace_na(UP, valencia.lithics.rf.probquantiles[1]),
EPI = replace_na(EPI, valencia.lithics.rf.probquantiles[1]),
MESO = replace_na(MESO, valencia.lithics.rf.probquantiles[1]),
ENEOL = replace_na(ENEOL, valencia.lithics.rf.probquantiles[1]),
LNEOL = replace_na(LNEOL, valencia.lithics.rf.probquantiles[1])) %>%
# occupational ubiquity
rename(MP_ubiq = MP,
UP_ubiq = UP,
EPI_ubiq = EPI,
MESO_ubiq = MESO,
ENEOL_ubiq = ENEOL,
LNEOL_ubiq = LNEOL) %>%
# calculate occupational intensity
mutate(MP_int = MP_ubiq*density.km2/800,
UP_int = UP_ubiq*density.km2/250,
EPI_int = EPI_ubiq*density.km2/40,
MESO_int = MESO_ubiq*density.km2/35,
ENEOL_int = ENEOL_ubiq*density.km2/25,
LNEOL_int = LNEOL_ubiq*density.km2/12)
# create file to graph results of patches with diagnostic lithics
medland.survey.rf.graph <- medland.survey.rf.graph %>%
dplyr::filter(!is.na(predicted.period)) %>%
pivot_longer(cols = 10:15,
names_to = "period",
values_to = "ubiquity") %>%
mutate(period = factor(period,
levels = c("MP","UP","EPI","MESO","ENEOL","MNEOL","LNEOL")),
age = case_when(period == "MP" ~ 80000,
period == "UP" ~ 25000,
period == "EPI" ~ 13000,
period == "MESO" ~ 9000,
period == "ENEOL" ~ 6000,
period == "LNEOL" ~ 4000),
predicted.age = case_when(
predicted.period == "MP" ~ 80000,
predicted.period == "UP" ~ 25000,
predicted.period == "EPI" ~ 13000,
predicted.period == "MESO" ~ 9000,
predicted.period == "ENEOL" ~ 6000,
predicted.period == "LNEOL" ~ 4000),
duration.centuries = case_when(
period == "MP" ~ 800,
period == "UP" ~ 250,
period == "EPI" ~ 40,
period == "MESO" ~ 35,
period == "ENEOL" ~ 25,
period == "LNEOL" ~ 12))Optional Graph: occupational ubiquity for all survey patches with diagnostic lithics
medland.survey.rf.graph %>%
ggplot(aes(x=period, y=ubiquity)) +
geom_rect(aes(fill = study.area),xmin = -Inf,xmax = Inf, ymin = -Inf,ymax = Inf,alpha = 0.1) +
geom_line(group=1) +
#geom_vline(aes(xintercept = predicted.period), color="blue") +
labs(title="Random Forest Age Estmates for Medland Survey Data",
subtitle="Time Periods for Each Assemblage, Colored by Study Area",
x="predicted time period\n(blue line indicates maximum predicted probability)") +
facet_wrap(~ assemblage + ID) +
theme_bw(base_size = 16) +
theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1))
Optional Graph: land use intensity for all survey patches with diagnostic lithics
medland.survey.rf.graph %>%
ggplot(aes(x=period, y=ubiquity*density.km2/duration.centuries)) +
geom_rect(aes(fill = study.area),xmin = -Inf,xmax = Inf, ymin = -Inf,ymax = Inf,alpha = 0.1) +
geom_line(group=1) +
#geom_vline(aes(xintercept = predicted.period), color="blue") +
labs(title="Land Use Intensity Estmates for Medland Survey Data",
subtitle="Time Periods for Each Assemblage, Colored by Study Area",
x="predicted time period\n(blue line indicates maximum predicted probability)",
y="estimated artifact accumulation rate\n(artifacts/km2/century)") +
facet_wrap(~ assemblage + ID) +
theme_bw(base_size = 16) +
theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1))
Figure 6a
medland.survey.rf.graph %>%
dplyr::filter(ID==238 | ID==960) %>%
ggplot(aes(x=period, y=ubiquity)) +
geom_line(group=1, size=2) +
labs(title="Occupational Ubiquity for 2 Survey Patches",
subtitle="Random Forest Age Estmates",
x="predicted time period",
y="occupational ubiquity") +
facet_wrap(~ assemblage) +
theme_bw(base_size = 20) +
theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1))
Figure 6b
medland.survey.rf.graph %>%
dplyr::filter(ID==238 | ID==960) %>%
ggplot(aes(x=period, y=ubiquity*density.km2/duration.centuries)) +
geom_line(group=1, size=2) +
labs(title="Land Use Intensity for 2 Survey Patches",
subtitle="Random Forest Age Estmates",
x="predicted time period",
y="artifacts / km^2 /century") +
facet_wrap(~ assemblage) +
theme_bw(base_size = 20) +
theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1),
axis.title.y = element_markdown())
Save random forest dating file to csv for mappoing in GIS
write_csv(medland.survey.rf.out, "medland_survey_rf.csv")Aggregate summaries of ubiquity and intensity for each valley
Calcuate area surveyed totals for each valley
totals.surveyed <- medland_survey2014_2017 %>%
select(study.area, area.sqm) %>%
group_by(study.area) %>%
summarise(total.area.sqm=sum(area.sqm))Figure 10a: aggregate ubiquity for each valley
left_join(medland.survey.rf.graph, totals.surveyed) %>%
dplyr::filter(ubiquity > quantile(medland.survey.rf.graph$ubiquity,
probs = .5)) %>%
group_by(study.area, period) %>%
summarize(area.sum = sum(area.sqm),
total.area.sqm = first(total.area.sqm)) %>%
select(study.area, period, area.sum, total.area.sqm) %>%
mutate(pct.coverage = area.sum/total.area.sqm) %>%
ungroup() %>%
add_row(study.area = "Canal de Navarrés",
period = "EPI",
area.sum = 0,
total.area.sqm = 4107582,
pct.coverage = 0) %>%
add_row(study.area = "Cocina/Catadau",
period = "EPI",
area.sum = 0,
total.area.sqm = 1095683,
pct.coverage = 0) %>%
add_row(study.area = "Cocina/Catadau",
period = "MESO",
area.sum = 0,
total.area.sqm = 1095683,
pct.coverage = 0) %>%
add_row(study.area = "Hoya de Buñol",
period = "EPI",
area.sum = 0,
total.area.sqm = 1551377,
pct.coverage = 0) %>%
# Needed for upper quartile graphing
# add_row(study.area = "Canal de Navarrés",
# period = "MP",
# area.sum = 0,
# total.area.sqm = 4107582,
# pct.coverage = 0) %>%
# add_row(study.area = "Canal de Navarrés",
# period = "MESO",
# area.sum = 0,
# total.area.sqm = 4107582,
# pct.coverage = 0) %>%
# add_row(study.area = "Canal de Navarrés",
# period = "LNEOL",
# area.sum = 0,
# total.area.sqm = 4107582,
# pct.coverage = 0) %>%
# add_row(study.area = "Cocina/Catadau",
# period = "MP",
# area.sum = 0,
# total.area.sqm = 1095683,
# pct.coverage = 0) %>%
# add_row(study.area = "Cocina/Catadau",
# period = "LNEOL",
# area.sum = 0,
# total.area.sqm = 1095683,
# pct.coverage = 0) %>%
# add_row(study.area = "Hoya de Buñol",
# period = "MP",
# area.sum = 0,
# total.area.sqm = 1551377,
# pct.coverage = 0) %>%
# add_row(study.area = "Hoya de Buñol",
# period = "MESO",
# area.sum = 0,
# total.area.sqm = 1551377,
# pct.coverage = 0) %>%
# add_row(study.area = "Hoya de Buñol",
# period = "LNEOL",
# area.sum = 0,
# total.area.sqm = 1551377,
# pct.coverage = 0) %>%
ggplot(aes(x=factor(period, levels =
c('MP','UP','EPI','MESO','ENEOL','LNEOL')),
y=pct.coverage, group=1)) +
geom_line(lwd=1.5) +
facet_wrap(~study.area, ncol = 1) +
labs(title='Aggregate Occupational Ubiquity',
subtitle='Proportion of study area surveyed\nwith ubiquity above the median',
x='period',
y='proportion') +
theme_bw(base_size = 16)Joining, by = "study.area"`summarise()` has grouped output by 'study.area'. You can override using the `.groups` argument.
Figure 10b: aggregate intensity for each valley
left_join(medland.survey.rf.graph, totals.surveyed) %>%
mutate(intensity = ubiquity*density.km2/duration.centuries) %>%
dplyr::filter(intensity >= quantile(medland.survey.rf.graph$ubiquity*medland.survey.rf.graph$density.km2/medland.survey.rf.graph$duration.centuries, probs =.5)) %>%
group_by(study.area, period) %>%
summarize(area.sum = sum(area.sqm),
total.area.sqm = first(total.area.sqm),
density.km2 = first(density.km2)) %>%
select(study.area, period, area.sum, total.area.sqm) %>%
mutate(pct.coverage = area.sum/total.area.sqm) %>%
ungroup %>%
add_row(study.area = "Canal de Navarrés",
period = "MP",
area.sum = 0,
total.area.sqm = 4107582,
pct.coverage = 0) %>%
add_row(study.area = "Hoya de Buñol",
period = "MP",
area.sum = 0,
total.area.sqm = 1551377,
pct.coverage = 0) %>%
# Needed for optional upper quartile graphing
# add_row(study.area = "Hoya de Buñol",
# period = "UP",
# area.sum = 0,
# total.area.sqm = 1551377,
# pct.coverage = 0) %>%
ggplot(aes(x=factor(period, levels =
c('MP','UP','EPI','MESO','ENEOL','LNEOL')),
y=pct.coverage, group=1)) +
geom_line(lwd=1.5) +
facet_wrap(~study.area, ncol = 1) +
labs(title='Aggregate Land Use Intensity',
subtitle='Proportion of study area surveyed\nwith intensity above the median',
x='period',
y='proportion') +
theme_bw(base_size = 16)Joining, by = "study.area"`summarise()` has grouped output by 'study.area'. You can override using the `.groups` argument.
Figure 11: SPD Analyses of Prehistoric Demography
Prepare Data
Only use dates with COV ≤ 0.05
C14_SE_Iberia_all <- C14_SE_Iberia_all %>% dplyr::filter(C14.CV<0.05 & C14.SD>0)Calibrate Dates with BChron
all.dates.calibrated <- with(C14_SE_Iberia_all, BchronCalibrate(ages = C14.mean, ageSds = C14.SD, calCurves = calib.curve, positions = site))
C14_SE_Iberia_all$BP.cal.median <- sapply(1:length(all.dates.calibrated), function(x) round(median(all.dates.calibrated[[x]]$ageGrid)))
Bin Dates
C14_SE_Iberia_all.bins <- C14_SE_Iberia_all %>%
with(., binPrep(site, C14.mean, 100))Model Test
C14_SE_Iberia_all.modeltest <- C14_SE_Iberia_all %>%
with(., calibrate(x=C14.mean, errors=C14.SD, calCurves = calib.curve, normalised=TRUE, calMatrix=FALSE)) %>%
modelTest(.,
errors = C14_SE_Iberia_all$C14.SD,
timeRange = c(35000,3000),
runm = 500,
model="exponential",
datenormalised=TRUE,
nsim = 200,
ncores = ncores,
method = 'calsample',
bins = C14_SE_Iberia_all.bins)[1] "Calibrating radiocarbon ages..."
|
| | 0%
|
|= | 1%
|
|= | 2%
|
|== | 2%
|
|== | 3%
|
|=== | 3%
|
|=== | 4%
|
|==== | 4%
|
|==== | 5%
|
|===== | 5%
|
|===== | 6%
|
|====== | 6%
|
|====== | 7%
|
|======= | 7%
|
|======= | 8%
|
|======== | 8%
|
|======== | 9%
|
|========= | 9%
|
|========= | 10%
|
|========== | 10%
|
|========== | 11%
|
|=========== | 11%
|
|=========== | 12%
|
|============ | 12%
|
|============ | 13%
|
|============= | 13%
|
|============= | 14%
|
|============== | 15%
|
|============== | 16%
|
|=============== | 16%
|
|=============== | 17%
|
|================ | 17%
|
|================ | 18%
|
|================= | 18%
|
|================= | 19%
|
|================== | 19%
|
|================== | 20%
|
|=================== | 20%
|
|=================== | 21%
|
|==================== | 21%
|
|==================== | 22%
|
|===================== | 22%
|
|===================== | 23%
|
|====================== | 23%
|
|====================== | 24%
|
|======================= | 24%
|
|======================= | 25%
|
|======================== | 25%
|
|======================== | 26%
|
|========================= | 26%
|
|========================= | 27%
|
|========================== | 27%
|
|========================== | 28%
|
|=========================== | 29%
|
|=========================== | 30%
|
|============================ | 30%
|
|============================ | 31%
|
|============================= | 31%
|
|============================= | 32%
|
|============================== | 32%
|
|============================== | 33%
|
|=============================== | 33%
|
|=============================== | 34%
|
|================================ | 34%
|
|================================ | 35%
|
|================================= | 35%
|
|================================= | 36%
|
|================================== | 36%
|
|================================== | 37%
|
|=================================== | 37%
|
|=================================== | 38%
|
|==================================== | 38%
|
|==================================== | 39%
|
|===================================== | 39%
|
|===================================== | 40%
|
|====================================== | 40%
|
|====================================== | 41%
|
|======================================= | 41%
|
|======================================= | 42%
|
|======================================== | 43%
|
|======================================== | 44%
|
|========================================= | 44%
|
|========================================= | 45%
|
|========================================== | 45%
|
|========================================== | 46%
|
|=========================================== | 46%
|
|=========================================== | 47%
|
|============================================ | 47%
|
|============================================ | 48%
|
|============================================= | 48%
|
|============================================= | 49%
|
|============================================== | 49%
|
|============================================== | 50%
|
|=============================================== | 50%
|
|=============================================== | 51%
|
|================================================ | 51%
|
|================================================ | 52%
|
|================================================= | 52%
|
|================================================= | 53%
|
|================================================== | 53%
|
|================================================== | 54%
|
|=================================================== | 54%
|
|=================================================== | 55%
|
|==================================================== | 55%
|
|==================================================== | 56%
|
|===================================================== | 56%
|
|===================================================== | 57%
|
|====================================================== | 58%
|
|====================================================== | 59%
|
|======================================================= | 59%
|
|======================================================= | 60%
|
|======================================================== | 60%
|
|======================================================== | 61%
|
|========================================================= | 61%
|
|========================================================= | 62%
|
|========================================================== | 62%
|
|========================================================== | 63%
|
|=========================================================== | 63%
|
|=========================================================== | 64%
|
|============================================================ | 64%
|
|============================================================ | 65%
|
|============================================================= | 65%
|
|============================================================= | 66%
|
|============================================================== | 66%
|
|============================================================== | 67%
|
|=============================================================== | 67%
|
|=============================================================== | 68%
|
|================================================================ | 68%
|
|================================================================ | 69%
|
|================================================================= | 69%
|
|================================================================= | 70%
|
|================================================================== | 70%
|
|================================================================== | 71%
|
|=================================================================== | 72%
|
|=================================================================== | 73%
|
|==================================================================== | 73%
|
|==================================================================== | 74%
|
|===================================================================== | 74%
|
|===================================================================== | 75%
|
|====================================================================== | 75%
|
|====================================================================== | 76%
|
|======================================================================= | 76%
|
|======================================================================= | 77%
|
|======================================================================== | 77%
|
|======================================================================== | 78%
|
|========================================================================= | 78%
|
|========================================================================= | 79%
|
|========================================================================== | 79%
|
|========================================================================== | 80%
|
|=========================================================================== | 80%
|
|=========================================================================== | 81%
|
|============================================================================ | 81%
|
|============================================================================ | 82%
|
|============================================================================= | 82%
|
|============================================================================= | 83%
|
|============================================================================== | 83%
|
|============================================================================== | 84%
|
|=============================================================================== | 84%
|
|=============================================================================== | 85%
|
|================================================================================ | 86%
|
|================================================================================ | 87%
|
|================================================================================= | 87%
|
|================================================================================= | 88%
|
|================================================================================== | 88%
|
|================================================================================== | 89%
|
|=================================================================================== | 89%
|
|=================================================================================== | 90%
|
|==================================================================================== | 90%
|
|==================================================================================== | 91%
|
|===================================================================================== | 91%
|
|===================================================================================== | 92%
|
|====================================================================================== | 92%
|
|====================================================================================== | 93%
|
|======================================================================================= | 93%
|
|======================================================================================= | 94%
|
|======================================================================================== | 94%
|
|======================================================================================== | 95%
|
|========================================================================================= | 95%
|
|========================================================================================= | 96%
|
|========================================================================================== | 96%
|
|========================================================================================== | 97%
|
|=========================================================================================== | 97%
|
|=========================================================================================== | 98%
|
|============================================================================================ | 98%
|
|============================================================================================ | 99%
|
|=============================================================================================| 100%
[1] "Done."── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster
── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──
✔ ggplot2 3.3.6 ✔ purrr 0.3.4
✔ tibble 3.1.7 ✔ dplyr 1.0.9
✔ tidyr 1.2.0 ✔ stringr 1.4.0
✔ readr 2.1.2 ✔ forcats 0.5.1
Attaching package: ‘snow’
The following objects are masked from ‘package:parallel’:
clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
clusterExport, clusterMap, clusterSplit, makeCluster, parApply,
parCapply, parLapply, parRapply, parSapply, splitIndices,
stopCluster[1] "Aggregating observed dates..."
[1] "Monte-Carlo test..."Plot SPD
par(mar=c(7,7,7,3))
plot(C14_SE_Iberia_all.modeltest, xlim = c(30000,4000), col.obs = 'black', lwd.obs = 5, drawaxes = F)
axis(1, cex.axis = 2, pos = -.01, at=(seq(30000,0, by=-5000)))axis(2, cex.axis = 2, pos = 30200)
mtext(side=1, line=5, "calibrated years BP", cex=2.5)mtext(side=2, line=4, "summed probability density", cex=2.5)
title(main=paste("Southern and Eastern Iberia SPD (N = ", nrow(C14_SE_Iberia_all), ")\n"), cex.main = 3)
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